Biometric databases can be useful for criminal investigation. A biometric database comprises records, each of which contains biometric sample data and associated biographic data. For example, when booking a criminal suspect, a police agency collects biometrics (e.g., fingerprints, iris, “mug shot”, DNA) and biographic information, such as a name, address, height, weight, etc. The collected biometrics and biographics are stored together as a database record. Further, a criminal investigation may recover latent biometric samples from a crime scene, which are submitted later to an automated biometric identification system (ABIS). The biometric identification system algorithmically compares the latent biometric samples to records in the biometric database in an attempt to ascertain the source individual or individuals of the latent biometric samples.
The mechanism of biometric matching, however, is probabilistic in nature, and does not attempt to yield 100% certainty in identification. For example, a biometric search typically does not yield a certain match, but rather a ranked candidate list, which highly trained forensic specialists then review manually. The biometric search results, however, can be inconclusive, particularly when the latent biometric sample is of poor quality.